Aim: The aim of this paper was to perform a meta-analysis of the literature to evaluate the accuracy of FDG-PET in recurrence detection in patients with ovarian cancer.
Methods: The search strategy consisted in identifying papers published between November 1972 and October 2003 indexed in MEDLINE and CANCERLITE. Inclusion criteria were publications that evaluated recurrence with FDG-PET in patients with ovarian carcinoma. Exclusion criteria were duplicated papers or those outdated by subsequent ones. The statistical analysis included 95 % confidence intervals (CI) of sensitivity (S), specificity (Sp) and natural logarithm of the odds ratio (Ln OR).
Results: Seventeen publications fulfilled the inclusion criteria and were analyzed. The S and Sp were 0.94 (95% CI, 0.93-0.96) and 0.65 (95% CI, 0.46-0.85), respectively. The Ln OR presented global homogeneity and significant values in > 75% of the studies.
Conclusion: According to our meta-analysis, FDG-PET has high S and intermediate Sp, with few false negative results. These preliminary findings suggest that FDG-PET may be an effective means of identifying patients with recurrent ovarian cancer.